xdiff/xprepare.c

   1 /*
   2  *  LibXDiff by Davide Libenzi ( File Differential Library )
   3  *  Copyright (C) 2003  Davide Libenzi
   4  *
   5  *  This library is free software; you can redistribute it and/or
   6  *  modify it under the terms of the GNU Lesser General Public
   7  *  License as published by the Free Software Foundation; either
   8  *  version 2.1 of the License, or (at your option) any later version.
   9  *
  10  *  This library is distributed in the hope that it will be useful,
  11  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  *  Lesser General Public License for more details.
  14  *
  15  *  You should have received a copy of the GNU Lesser General Public
  16  *  License along with this library; if not, write to the Free Software
  17  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  18  *
  19  *  Davide Libenzi <davidel@xmailserver.org>
  20  *
  21  */
  22
  23 #include "xinclude.h"
  24
  25
  26 #define XDL_KPDIS_RUN 4
  27 #define XDL_MAX_EQLIMIT 1024
  28 #define XDL_SIMSCAN_WINDOW 100
  29
  30
  31 typedef struct s_xdlclass {
  32         struct s_xdlclass *next;
  33         unsigned long ha;
  34         char const *line;
  35         long size;
  36         long idx;
  37 } xdlclass_t;
  38
  39 typedef struct s_xdlclassifier {
  40         unsigned int hbits;
  41         long hsize;
  42         xdlclass_t **rchash;
  43         chastore_t ncha;
  44         long count;
  45         long flags;
  46 } xdlclassifier_t;
  47
  48
  49
  50
  51 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags);
  52 static void xdl_free_classifier(xdlclassifier_t *cf);
  53 static int xdl_classify_record(xdlclassifier_t *cf, xrecord_t **rhash, unsigned int hbits,
  54                                xrecord_t *rec);
  55 static int xdl_prepare_ctx(mmfile_t *mf, long narec, xpparam_t const *xpp,
  56                            xdlclassifier_t *cf, xdfile_t *xdf);
  57 static void xdl_free_ctx(xdfile_t *xdf);
  58 static int xdl_clean_mmatch(char const *dis, long i, long s, long e);
  59 static int xdl_cleanup_records(xdfile_t *xdf1, xdfile_t *xdf2);
  60 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2);
  61 static int xdl_optimize_ctxs(xdfile_t *xdf1, xdfile_t *xdf2);
  62
  63
  64
  65
  66 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags) {
  67         long i;
  68
  69         cf->flags = flags;
  70
  71         cf->hbits = xdl_hashbits((unsigned int) size);
  72         cf->hsize = 1 << cf->hbits;
  73
  74         if (xdl_cha_init(&cf->ncha, sizeof(xdlclass_t), size / 4 + 1) < 0) {
  75
  76                 return -1;
  77         }
  78         if (!(cf->rchash = (xdlclass_t **) xdl_malloc(cf->hsize * sizeof(xdlclass_t *)))) {
  79
  80                 xdl_cha_free(&cf->ncha);
  81                 return -1;
  82         }
  83         for (i = 0; i < cf->hsize; i++)
  84                 cf->rchash[i] = NULL;
  85
  86         cf->count = 0;
  87
  88         return 0;
  89 }
  90
  91
  92 static void xdl_free_classifier(xdlclassifier_t *cf) {
  93
  94         xdl_free(cf->rchash);
  95         xdl_cha_free(&cf->ncha);
  96 }
  97
  98
  99 static int xdl_classify_record(xdlclassifier_t *cf, xrecord_t **rhash, unsigned int hbits,
 100                                xrecord_t *rec) {
 101         long hi;
 102         char const *line;
 103         xdlclass_t *rcrec;
 104
 105         line = rec->ptr;
 106         hi = (long) XDL_HASHLONG(rec->ha, cf->hbits);
 107         for (rcrec = cf->rchash[hi]; rcrec; rcrec = rcrec->next)
 108                 if (rcrec->ha == rec->ha &&
 109                                 xdl_recmatch(rcrec->line, rcrec->size,
 110                                         rec->ptr, rec->size, cf->flags))
 111                         break;
 112
 113         if (!rcrec) {
 114                 if (!(rcrec = xdl_cha_alloc(&cf->ncha))) {
 115
 116                         return -1;
 117                 }
 118                 rcrec->idx = cf->count++;
 119                 rcrec->line = line;
 120                 rcrec->size = rec->size;
 121                 rcrec->ha = rec->ha;
 122                 rcrec->next = cf->rchash[hi];
 123                 cf->rchash[hi] = rcrec;
 124         }
 125
 126         rec->ha = (unsigned long) rcrec->idx;
 127
 128         hi = (long) XDL_HASHLONG(rec->ha, hbits);
 129         rec->next = rhash[hi];
 130         rhash[hi] = rec;
 131
 132         return 0;
 133 }
 134
 135
 136 static int xdl_prepare_ctx(mmfile_t *mf, long narec, xpparam_t const *xpp,
 137                            xdlclassifier_t *cf, xdfile_t *xdf) {
 138         unsigned int hbits;
 139         long i, nrec, hsize, bsize;
 140         unsigned long hav;
 141         char const *blk, *cur, *top, *prev;
 142         xrecord_t *crec;
 143         xrecord_t **recs, **rrecs;
 144         xrecord_t **rhash;
 145         unsigned long *ha;
 146         char *rchg;
 147         long *rindex;
 148
 149         if (xdl_cha_init(&xdf->rcha, sizeof(xrecord_t), narec / 4 + 1) < 0) {
 150
 151                 return -1;
 152         }
 153         if (!(recs = (xrecord_t **) xdl_malloc(narec * sizeof(xrecord_t *)))) {
 154
 155                 xdl_cha_free(&xdf->rcha);
 156                 return -1;
 157         }
 158
 159         hbits = xdl_hashbits((unsigned int) narec);
 160         hsize = 1 << hbits;
 161         if (!(rhash = (xrecord_t **) xdl_malloc(hsize * sizeof(xrecord_t *)))) {
 162
 163                 xdl_free(recs);
 164                 xdl_cha_free(&xdf->rcha);
 165                 return -1;
 166         }
 167         for (i = 0; i < hsize; i++)
 168                 rhash[i] = NULL;
 169
 170         nrec = 0;
 171         if ((cur = blk = xdl_mmfile_first(mf, &bsize)) != NULL) {
 172                 for (top = blk + bsize;;) {
 173                         if (cur >= top) {
 174                                 if (!(cur = blk = xdl_mmfile_next(mf, &bsize)))
 175                                         break;
 176                                 top = blk + bsize;
 177                         }
 178                         prev = cur;
 179                         hav = xdl_hash_record(&cur, top, xpp->flags);
 180                         if (nrec >= narec) {
 181                                 narec *= 2;
 182                                 if (!(rrecs = (xrecord_t **) xdl_realloc(recs, narec * sizeof(xrecord_t *)))) {
 183
 184                                         xdl_free(rhash);
 185                                         xdl_free(recs);
 186                                         xdl_cha_free(&xdf->rcha);
 187                                         return -1;
 188                                 }
 189                                 recs = rrecs;
 190                         }
 191                         if (!(crec = xdl_cha_alloc(&xdf->rcha))) {
 192
 193                                 xdl_free(rhash);
 194                                 xdl_free(recs);
 195                                 xdl_cha_free(&xdf->rcha);
 196                                 return -1;
 197                         }
 198                         crec->ptr = prev;
 199                         crec->size = (long) (cur - prev);
 200                         crec->ha = hav;
 201                         recs[nrec++] = crec;
 202
 203                         if (xdl_classify_record(cf, rhash, hbits, crec) < 0) {
 204
 205                                 xdl_free(rhash);
 206                                 xdl_free(recs);
 207                                 xdl_cha_free(&xdf->rcha);
 208                                 return -1;
 209                         }
 210                 }
 211         }
 212
 213         if (!(rchg = (char *) xdl_malloc((nrec + 2) * sizeof(char)))) {
 214
 215                 xdl_free(rhash);
 216                 xdl_free(recs);
 217                 xdl_cha_free(&xdf->rcha);
 218                 return -1;
 219         }
 220         memset(rchg, 0, (nrec + 2) * sizeof(char));
 221
 222         if (!(rindex = (long *) xdl_malloc((nrec + 1) * sizeof(long)))) {
 223
 224                 xdl_free(rchg);
 225                 xdl_free(rhash);
 226                 xdl_free(recs);
 227                 xdl_cha_free(&xdf->rcha);
 228                 return -1;
 229         }
 230         if (!(ha = (unsigned long *) xdl_malloc((nrec + 1) * sizeof(unsigned long)))) {
 231
 232                 xdl_free(rindex);
 233                 xdl_free(rchg);
 234                 xdl_free(rhash);
 235                 xdl_free(recs);
 236                 xdl_cha_free(&xdf->rcha);
 237                 return -1;
 238         }
 239
 240         xdf->nrec = nrec;
 241         xdf->recs = recs;
 242         xdf->hbits = hbits;
 243         xdf->rhash = rhash;
 244         xdf->rchg = rchg + 1;
 245         xdf->rindex = rindex;
 246         xdf->nreff = 0;
 247         xdf->ha = ha;
 248         xdf->dstart = 0;
 249         xdf->dend = nrec - 1;
 250
 251         return 0;
 252 }
 253
 254
 255 static void xdl_free_ctx(xdfile_t *xdf) {
 256
 257         xdl_free(xdf->rhash);
 258         xdl_free(xdf->rindex);
 259         xdl_free(xdf->rchg - 1);
 260         xdl_free(xdf->ha);
 261         xdl_free(xdf->recs);
 262         xdl_cha_free(&xdf->rcha);
 263 }
 264
 265
 266 int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp,
 267                     xdfenv_t *xe) {
 268         long enl1, enl2;
 269         xdlclassifier_t cf;
 270
 271         enl1 = xdl_guess_lines(mf1) + 1;
 272         enl2 = xdl_guess_lines(mf2) + 1;
 273
 274         if (xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0) {
 275
 276                 return -1;
 277         }
 278
 279         if (xdl_prepare_ctx(mf1, enl1, xpp, &cf, &xe->xdf1) < 0) {
 280
 281                 xdl_free_classifier(&cf);
 282                 return -1;
 283         }
 284         if (xdl_prepare_ctx(mf2, enl2, xpp, &cf, &xe->xdf2) < 0) {
 285
 286                 xdl_free_ctx(&xe->xdf1);
 287                 xdl_free_classifier(&cf);
 288                 return -1;
 289         }
 290
 291         xdl_free_classifier(&cf);
 292
 293         if (xdl_optimize_ctxs(&xe->xdf1, &xe->xdf2) < 0) {
 294
 295                 xdl_free_ctx(&xe->xdf2);
 296                 xdl_free_ctx(&xe->xdf1);
 297                 return -1;
 298         }
 299
 300         return 0;
 301 }
 302
 303
 304 void xdl_free_env(xdfenv_t *xe) {
 305
 306         xdl_free_ctx(&xe->xdf2);
 307         xdl_free_ctx(&xe->xdf1);
 308 }
 309
 310
 311 static int xdl_clean_mmatch(char const *dis, long i, long s, long e) {
 312         long r, rdis0, rpdis0, rdis1, rpdis1;
 313
 314         /*
 315          * Limits the window the is examined during the similar-lines
 316          * scan. The loops below stops when dis[i - r] == 1 (line that
 317          * has no match), but there are corner cases where the loop
 318          * proceed all the way to the extremities by causing huge
 319          * performance penalties in case of big files.
 320          */
 321         if (i - s > XDL_SIMSCAN_WINDOW)
 322                 s = i - XDL_SIMSCAN_WINDOW;
 323         if (e - i > XDL_SIMSCAN_WINDOW)
 324                 e = i + XDL_SIMSCAN_WINDOW;
 325
 326         /*
 327          * Scans the lines before 'i' to find a run of lines that either
 328          * have no match (dis[j] == 0) or have multiple matches (dis[j] > 1).
 329          * Note that we always call this function with dis[i] > 1, so the
 330          * current line (i) is already a multimatch line.
 331          */
 332         for (r = 1, rdis0 = 0, rpdis0 = 1; (i - r) >= s; r++) {
 333                 if (!dis[i - r])
 334                         rdis0++;
 335                 else if (dis[i - r] == 2)
 336                         rpdis0++;
 337                 else
 338                         break;
 339         }
 340         /*
 341          * If the run before the line 'i' found only multimatch lines, we
 342          * return 0 and hence we don't make the current line (i) discarded.
 343          * We want to discard multimatch lines only when they appear in the
 344          * middle of runs with nomatch lines (dis[j] == 0).
 345          */
 346         if (rdis0 == 0)
 347                 return 0;
 348         for (r = 1, rdis1 = 0, rpdis1 = 1; (i + r) <= e; r++) {
 349                 if (!dis[i + r])
 350                         rdis1++;
 351                 else if (dis[i + r] == 2)
 352                         rpdis1++;
 353                 else
 354                         break;
 355         }
 356         /*
 357          * If the run after the line 'i' found only multimatch lines, we
 358          * return 0 and hence we don't make the current line (i) discarded.
 359          */
 360         if (rdis1 == 0)
 361                 return 0;
 362         rdis1 += rdis0;
 363         rpdis1 += rpdis0;
 364
 365         return rpdis1 * XDL_KPDIS_RUN < (rpdis1 + rdis1);
 366 }
 367
 368
 369 /*
 370  * Try to reduce the problem complexity, discard records that have no
 371  * matches on the other file. Also, lines that have multiple matches
 372  * might be potentially discarded if they happear in a run of discardable.
 373  */
 374 static int xdl_cleanup_records(xdfile_t *xdf1, xdfile_t *xdf2) {
 375         long i, nm, rhi, nreff, mlim;
 376         unsigned long hav;
 377         xrecord_t **recs;
 378         xrecord_t *rec;
 379         char *dis, *dis1, *dis2;
 380
 381         if (!(dis = (char *) xdl_malloc(xdf1->nrec + xdf2->nrec + 2))) {
 382
 383                 return -1;
 384         }
 385         memset(dis, 0, xdf1->nrec + xdf2->nrec + 2);
 386         dis1 = dis;
 387         dis2 = dis1 + xdf1->nrec + 1;
 388
 389         if ((mlim = xdl_bogosqrt(xdf1->nrec)) > XDL_MAX_EQLIMIT)
 390                 mlim = XDL_MAX_EQLIMIT;
 391         for (i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart]; i <= xdf1->dend; i++, recs++) {
 392                 hav = (*recs)->ha;
 393                 rhi = (long) XDL_HASHLONG(hav, xdf2->hbits);
 394                 for (nm = 0, rec = xdf2->rhash[rhi]; rec; rec = rec->next)
 395                         if (rec->ha == hav && ++nm == mlim)
 396                                 break;
 397                 dis1[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 398         }
 399
 400         if ((mlim = xdl_bogosqrt(xdf2->nrec)) > XDL_MAX_EQLIMIT)
 401                 mlim = XDL_MAX_EQLIMIT;
 402         for (i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart]; i <= xdf2->dend; i++, recs++) {
 403                 hav = (*recs)->ha;
 404                 rhi = (long) XDL_HASHLONG(hav, xdf1->hbits);
 405                 for (nm = 0, rec = xdf1->rhash[rhi]; rec; rec = rec->next)
 406                         if (rec->ha == hav && ++nm == mlim)
 407                                 break;
 408                 dis2[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 409         }
 410
 411         for (nreff = 0, i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart];
 412              i <= xdf1->dend; i++, recs++) {
 413                 if (dis1[i] == 1 ||
 414                     (dis1[i] == 2 && !xdl_clean_mmatch(dis1, i, xdf1->dstart, xdf1->dend))) {
 415                         xdf1->rindex[nreff] = i;
 416                         xdf1->ha[nreff] = (*recs)->ha;
 417                         nreff++;
 418                 } else
 419                         xdf1->rchg[i] = 1;
 420         }
 421         xdf1->nreff = nreff;
 422
 423         for (nreff = 0, i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart];
 424              i <= xdf2->dend; i++, recs++) {
 425                 if (dis2[i] == 1 ||
 426                     (dis2[i] == 2 && !xdl_clean_mmatch(dis2, i, xdf2->dstart, xdf2->dend))) {
 427                         xdf2->rindex[nreff] = i;
 428                         xdf2->ha[nreff] = (*recs)->ha;
 429                         nreff++;
 430                 } else
 431                         xdf2->rchg[i] = 1;
 432         }
 433         xdf2->nreff = nreff;
 434
 435         xdl_free(dis);
 436
 437         return 0;
 438 }
 439
 440
 441 /*
 442  * Early trim initial and terminal matching records.
 443  */
 444 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
 445         long i, lim;
 446         xrecord_t **recs1, **recs2;
 447
 448         recs1 = xdf1->recs;
 449         recs2 = xdf2->recs;
 450         for (i = 0, lim = XDL_MIN(xdf1->nrec, xdf2->nrec); i < lim;
 451              i++, recs1++, recs2++)
 452                 if ((*recs1)->ha != (*recs2)->ha)
 453                         break;
 454
 455         xdf1->dstart = xdf2->dstart = i;
 456
 457         recs1 = xdf1->recs + xdf1->nrec - 1;
 458         recs2 = xdf2->recs + xdf2->nrec - 1;
 459         for (lim -= i, i = 0; i < lim; i++, recs1--, recs2--)
 460                 if ((*recs1)->ha != (*recs2)->ha)
 461                         break;
 462
 463         xdf1->dend = xdf1->nrec - i - 1;
 464         xdf2->dend = xdf2->nrec - i - 1;
 465
 466         return 0;
 467 }
 468
 469
 470 static int xdl_optimize_ctxs(xdfile_t *xdf1, xdfile_t *xdf2) {
 471
 472         if (xdl_trim_ends(xdf1, xdf2) < 0 ||
 473             xdl_cleanup_records(xdf1, xdf2) < 0) {
 474
 475                 return -1;
 476         }
 477
 478         return 0;
 479 }